JPS61273682A - Character correction system for optical character reader - Google Patents

Abstract

PURPOSE:To perform a rejecting correction, a misreading correction etc. easily and quickly with high efficiency, by displaying a simple candidate character read out of a simple candidate register file in a character correction mode as a selection subject. CONSTITUTION:The data obtained by reading a form through an OCR main body 10 is sent to a control part via a reception part 21 for checking the presence/absence of rejection. When a rejecting action is detected, the simple candidates are sent out of a file 24 to a display mechanism 22 for display together with the form image data, a rejection image pattern, the recognition candidates, etc. Here an operator searches a subject character out of those recognition candidates based on the rejection image pattern for input of the subject character. At the same time, the operator selects a character out of the simple candidates and displays it.

Description

DETAILED DESCRIPTION OF THE INVENTION [Technical Field of the Invention] The present invention relates to an optical character reading device, and particularly to a character correction method for an optical character reading device, which is characterized by a candidate character selection mechanism during correction. O [Technical background of the invention and its problems] Conventionally, character correction (reject characters or misreading) in an optical character reading device (hereinafter referred to as OCR) is performed using a quantized image displayed on a display screen or an actual document. I checked and found that the key was performing some key operations.

At this time, if the read characters to be corrected are alphanumeric characters/kana characters/symbols, etc., touch 1 or 2.
Corrections can be made with a simple key operation, such as a touch, and there is no particularly large change in character. However, if the character to be corrected is a kanji, the correction can be performed using, for example, onkun specification input, kana-kanji conversion input, Romaji conversion input or Kanji 1187
All corrections, such as entering a kuten code, must be performed using correction means that require many key operations.

That is, for example, in the case of correction using the sound lesson specified input method, (
[il specified key + (sound or lesson continuation data: number of key operations α
) + Data selection Jukan γ key] key operation of touching "4 + α" is required, and in the case of correction using the kana-kanji conversion input method, C 匪 5] Han designation key + (Reading data; key operation ``3+'' by operating the number β)
In addition, in the case of correction using the Kanji JIS code input method, the key operation of touching "6" is required by operating [B34 designated key 10 ml key x 4 10 n key] Is required.

In this way, the kanji code should be corrected at least 5 to 6 times,
At most, it would be more than 10 tatsu'f-, making the operation complicated. In addition, in recent years there has been a correction method using candidate characters from OCR recognition results, but if there is no correction data in the list, many keys must be manually operated as described above, making the operation complicated. There were drawbacks.

[Purpose of the invention]

The present invention was made in view of the above-mentioned circumstances.
When correcting characters in the event of misreading, etc., candidate characters registered in advance can be displayed as selection targets in consideration of the possibility of rejection or misreading, making it easy and quick to correct characters. An object of the present invention is to provide a character correction method for an optical character reading device.

[Summary of the invention]

The present invention provides an optical character reading device that includes a simple candidate registration file that stores simple candidate characters in consideration of the possibility of rejection or misreading, and a simple candidate registration mechanism that registers the simple candidate characters in this file. However, when modifying characters, the simple candidate characters that are successively added from the simple candidate registration file are displayed as one selection.
Reject corrections, misreading corrections, etc. can be quickly, easily, and efficiently executed.

[Embodiments of the invention]

An embodiment of the present invention will be described below with reference to the drawings.

FIG. 1 is a flowchart showing the main components of an apparatus according to an embodiment of the present invention.

In the figure, 10 is an OCR main body that controls OCR transport and recognition. 20 to 24 each constitute a component of a device that processes data read by the OCR main body 10;
22 is a data receiving unit that receives one form's worth of data read by the OCR main body 10; 22 is a display mechanism that displays a form image, image arm turn, recognition candidates, simple candidates, etc.; 23;
2 is a simple candidate registration mechanism that registers simple candidates in a specific file, and 24 is a simple candidate registration file that stores simple candidates.

FIG. 2 shows a display example when a reject is corrected, and FIG. 3 shows a display example when a misreading is corrected. Here, in addition to the recognition candidate CC, a simple candidate C8 is also displayed.

The above simple sentence [C8 is a key for characters that are easy to reject, characters that are easy to misread, etc.? - This is a group of candidate characters that are input by key operation and stored in advance in the simple candidate registration file 24 under the registration control of the simple candidate registration mechanism 23. Then, the simple suggestion 1cs stored in the simple suggestion Fi 7 isle 24 is read out when a rejected character or misread character is corrected, and sent to the display mechanism 22 to be displayed as shown in the figure.

The flow shown in FIG. 1 will be explained. First, scan the form using the OCR main unit 10.
Read with After the read data is received by the data receiving section 21, it is sent to a control section (not shown), and the presence or absence of a reject is checked. If there is a rejection, the display mechanism 22
As shown in FIG. 2, the form image data person, reject image pattern B, recognition candidate CC, etc. are displayed on the screen, and at the same time, the above-mentioned simple candidates are also displayed.

When a rejection occurs, the operator first looks at the reject image pattern and searches for the target character among the recognition candidates (candidates recognized by OCRC self-recognition chic). If the target character is found in the candidates, the character can be input by inputting its display designation number. If there are no candidates, generally the kanji is input by onkun code input, kana-kanji conversion, code input, etc., but here we use simple candidate C, which is registered in advance.
It is also possible to select from S. The simple candidates are a group of characters that are stored in the registration file 24, such as letters and characters that are easily misread (those that are input frequently from the keys), and the necessary characters are extracted and displayed. At this time, it is possible to select the necessary kanji by attaching various conditions (for example, depending on the business) as to whether the kanji is a necessary character or not. It is also assumed that editing such as adding data to the registered file 24, deleting data, and changing the display order can be performed.

Also, when correcting misread characters, simple candidate C8 is the third
It will be displayed as shown in the figure.

With the above-described simple candidate registration and display functions, when modifying characters, it is now possible to manually enter any character with about 3 touches instead of the conventional 3 to 10 key touches. In addition, kanji codes that suit the user can be freely registered and displayed when making corrections.

The above-mentioned simple candidate registration, display, and selection mechanism is based on OCR.
It can be applied not only to the device but also to other kanji key manual machines.

〔Effect of the invention〕

As described above, according to the present invention, in an optical character reading device, there is a simple candidate registration file that stores simple candidate characters considering the possibility of rejection or misreading, and the simple candidate characters are registered in this file. By having a simple candidate registration mechanism, and displaying the simple candidate characters successively generated from the above-mentioned simple candidate registration file as selection targets when correcting characters, reject corrections, misreading corrections, etc. can be performed quickly, easily, and efficiently. can do.

[Brief explanation of the drawing]

FIG. 1 is a block diagram showing the components in a flow section in one embodiment of the present invention, and FIGS. 2 and 3 are diagrams showing display examples when correcting characters in the above embodiment, respectively. be. lO...OCR main body, 21...Data receiving section, 22
... Display mechanism, 23... Simple candidate registration mechanism, 24.
...Simple candidate registration file, CC...Recognition candidate, CS
...Simple candidate.

Claims (1)

[Claims] In an optical character reading device, there is provided a simple candidate character storage mechanism for storing simple candidate characters considering the possibility of rejection or misreading, and a simple candidate character registration mechanism for registering the simple candidate characters in the simple candidate character storage mechanism. , an access mechanism for reading and controlling the simple candidate character storage mechanism when modifying characters; a display mechanism for displaying the simple candidate characters read from the simple candidate character storage mechanism by this access mechanism; and a simple candidate displayed on the display mechanism. 1. A character correction method for an optical character reading device, comprising a character specifying means for selecting an arbitrary character among characters, and displaying a simple candidate character as one selection target when correcting a character.